Targets to prevent prolonged length of stay after endovascular aortic repair

Abstract

Objective

Endovascular aneurysm repair (EVAR) is a commonly performed vascular operation. Yet, postoperative length of stay (LOS) varies greatly, even within institutions. The present study reviewed the morbidity, mortality, and the financial effect of increased LOS to establish modifiable factors associated with prolonged hospital LOS, with the goal of improving quality.

Methods

The Society for Vascular Surgery Vascular Quality Initiative database was used to identify all patients under-going primary, elective EVAR at a single institution between January 1, 2011, and May 28, 2014. Preoperative patient characteristics, intraoperative details, postoperative factors, long-term outcomes, and cost data were reviewed using an Institutional Review Board-approved prospectively collected database. Multivariate analysis was used to determine statistical difference between patients with LOS ≤2 days and >2 days.

Results

Complete 30-day variable and cost data were available for 138 patients with an average follow-up of 12 months; of these, 46 (33%) had a LOS >2 days. Variables determined to be statistically significant predictors of prolonged LOS included aneurysm diameter (P = .03), American Society of Anesthesiologists Physical Status Classification score (P < .001), thromboembolectomy (P = .01), and increased postoperative cardiac (P < .001) and renal (P = .01) complications. Specifically, modifiable risk factors that contributed to increased LOS included performance of a concomitant procedure (P < .001), increased volume of iodinated contrast (P = .05), increased volume of intraoperative crystalloid (P = .05), placement in an intensive care unit (P < .001), return to the operating room (P < .001), and the use of vasoactive medications (P < .001). Hospital charges ($102,000 ± $41,000 vs $180,000 ± $73,000; P = .01) and costs ($27,000 ± $10,000 vs $45,000 ± $19,000 P = .01) were significantly higher in patients with prolonged LOS; however, there was no difference in physician charges ($8000 ± $5700 vs $12,000 ± $12,000; P = .09). Increased LOS after EVAR was associated with an increase in mortality at 1 month (0% vs 4% P = .05) and 12 months (3% vs 13% P = .03).

Conclusions

This study highlights several modifiable risk factors leading to increased LOS after EVAR, including performance of concomitant procedures, admission to the intensive care unit, and postoperative renal and cardiac complications. Further, increased LOS was associated with increased charges, costs, morbidity, and mortality after EVAR. This study highlights specific areas of focus for decreasing LOS after EVAR and, in turn, improving quality in vascular surgery.

Health care costs continue to grow, with some projections anticipating expenditures to reach 20% of the United States gross domestic product by 2020.¹ As a result, national attention has focused on providing high-quality with concurrent cost-effective care. Reducing hospital length of stay (LOS) after routine procedures has been cited as a quality metric to reduce cost and hospital-acquired morbidity. In a world of value-based purchasing driven by third-party payers, these metrics, such as LOS, are more important than ever.

As endovascular techniques continue to improve, the use of endovascular aneurysm repair (EVAR) in the treatment of abdominal aortic aneurysm (AAA) has become the primary treatment paradigm during the past two decades.^2-8 However, these less invasive therapies are accompanied by significant procedure-related costs.^9,10 Specifically for EVAR, a hospital LOS >3 days has been associated with increased morbidity and mortality, according to a study by Cadili et al.² Each additional day in the hospital is estimated to increase the overall cost of the stay by 8% in some studies.¹¹

Perioperative care for many operations is becoming pathway driven and is highly standardized; however, these protocols vary significantly between hospitals and providers.¹² After the Virginia’s Vascular Study Group Regional meeting in 2014, our institution was noted to have a higher than expected observed-to-expected ratio for LOS in patients undergoing EVAR. This is attributed to our institution being the major referral center in our region with a high case complexity compared with surrounding centers. However, because EVAR is one of the most frequently performed vascular operations, we sought to identify opportunities for better preoperative and discharge planning, focusing on modifiable risk factors to shorten LOS and thus improve outcomes. We hypothesized that a number of preoperative, intraoperative, and postoperative factors could be identified and thus modified to reduce the incidence of prolonged LOS in our EVAR population.

METHODS

This study was approved by the University of Virginia Institutional Review Board (IRB No. 17900). The requirement for patient consent was waived.

Patients

All patients undergoing primary standard elective EVAR performed for infrarenal AAA at the University of Virginia Health System between January 1, 2011, and May 28, 2014, were reviewed. Patients undergoing graft modification (extensions, conversions to aortouniiliac), treatment of endoleaks or urgent/emergency repair were excluded. Data were obtained using the prospectively collected Society for Vascular Surgery Vascular Quality Initiative database. Preoperative comorbidities, operative variables, postoperative complications, and 30-day and 1-year mortality were analyzed to determine the difference between patients with postoperative LOS ≤2 days and those with LOS >2 days. Complete 30-day follow-up was available on all patients included.

Definitions

Congestive heart failure (CHF) is stratified as mild (slight limitation of physical activity, comfortable at rest, but ordinary physical activity results in fatigue, palpitation, or dyspnea), moderate (marked limitation of physical activity, comfortable at rest, but less than ordinary activity causes fatigue, palpitation, or dyspnea) and severe (unable to perform any physical activity without discomfort, symptoms of cardiac insufficiency at rest).

Coronary artery disease is similarly defined as stable angina (stable pattern or symptoms with or without antianginal medication) or unstable angina (new onset, increasing frequency, lasting >20 minutes or rest angina, or both).

Arterial injury is any injury requiring intervention or resulting in occlusion.

A symptomatic aneurysm is any EVAR performed ≤24 hours of pain or tenderness, without rupture. The ruptured classification requires computed tomography or angiography evidence of rupture.

Change in renal function is defined as a new increase in creatinine of 0.5 mg/dL.

Dysrhythmia is any new rhythm disturbance requiring treatment with medications or cardioversion. Myocardial infarction (MI) is classified as electrocardiogram changes or clinical evidence of MI in conjunction with any abnormality of cardiac biomarker consistent with infarction.

Respiratory complications include pneumonia with documented lobar infiltrate on a chest X-ray image and pure growth of a recognized pathogen.

Leg ischemia is defined as loss of previously palpable pulses, loss of previously present Doppler signals, and a decrease of >0.15 in the ankle-brachial index with blue toe or tissue loss. Bowel ischemia is diagnosed by colonoscopic evidence of ischemia, bloody stools in a patient who dies before colonoscopy or laparotomy, or presumptive diagnosis with conservative treatment.

Statistical analysis

Primary outcomes included preoperative comorbidities, operative variables, and postoperative complications. Secondary outcomes included morbidity or mortality at 1, 6, and 12 months in addition to hospital charges and cost as well as physician charges. Differences between the standard LOS (≤2 days) and prolonged LOS (>2 days) study groups were compared using the χ² test for categoric variables and with Mann-Whitney U test for continuous variables. P < .05 was used for statistical significance.

RESULTS

Patient characteristics and preoperative risk factors for increased LOS

Of the 138 patients undergoing EVAR during the past 3 years, 46 (33%) had a LOS >2 days. All patients included in the study had complete 30-day follow-up, with an average of 13-month follow-up in those with LOS ≤2 days and 12 months for those with LOS >2 days (P = .18). The average LOS for all patients was 2.7 ± 2.3 days. Table I summarizes preoperative characteristics for patients undergoing EVAR. White race (P = .03), American Society of Anesthesiologists (ASA) Physical Status Classification score (P < .001), iliac aneurysm (P = .01), severe CHF (P = .04), and preoperative angiotensin-converting enzyme inhibitor (ACE-I) therapy (P < .03) were the only variables to reach statistical significance for predicting prolonged LOS. All procedures were performed by four surgeons with varying experience at a high-volume academic medical center, with no statistically significant difference in LOS (P = .22) between surgeons.

Table I.

Preoperative patient characteristics of endovascular aneurysm repair (EVAR) groups by length of stay (LOS)

	LOS ≤2 days		LOS >2 days
Characteristics	No.	%	No.	%	χ 2	df	Pa
Male sex	78	85	39	85	0.00	1	1
White race	86	94	42	91	11.02	4	.03
Surgeon	44	48	25	54	2.30	3	.51
Medicare	74	80	40	87	0.91	1	.34
Medicaid	92	100	46	100	0.00	1	1
Private insurance	11	12	4	9	0.34	1	.56
Veterans Affairs	1	1	2	4	1.53	1	.22
Self-pay	6	7	0	0	3.14	1	.08
Never smoker	10	11	4	9	0.16	1	.69
Quit smoking	45	49	28	61	1.76	1	.19
Current smoker	37	40	14	30	1.26	1	.26
Hypertension	75	82	38	83	0.02	1	.88
Diabetes mellitus
None	70	76	37	80	0.33	1	.56
Diet controlled	8	9	3	7	0.20	1	.66
Medication controlled	9	10	5	11	0.04	1	.84
Insulin controlled	5	5	1	2	0.78	1	.38
Coronary artery disease
Stable	27	29	11	24	0.45	1	.50
Unstable	4	4	1	2	0.42	1	.52
Coronary artery bypass graft or stent
>5 years ago	17	19	11	24	0.56	1	.45
<5 years ago	11	12	5	11	0.04	1	.85
CHF
Mild	2	2	3	7	1.66	1	.20
Moderate	3	3	3	7	0.78	1	.38
Severe	0	0	2	4	4.06	1	.04
Left ventricular ejection fraction
<30%	4	4	3	7	0.30	1	.58
30%-50%	11	12	7	15	0.29	1	.59
Chronic obstructive pulmonary disease
Asymptomatic	4	4	3	7	0.30	1	.58
Medication	16	17	7	15	0.10	1	.75
Oxygen dependent	3	3	5	11	3.25	1	.07
Currently on dialysis	1	1	0	0	0.50	1	.48
ASA Physical Status Classification
2	9	10	0	0
3	71	77	25	54	20.43	3	<.001
4	12	13	21	46
Medication
β-blocker	28	30	11	24	0.64	1	.42
Aspirin	66	72	35	76	0.30	1	.59
Plavixb	12	13	3	7	1.35	1	.25
Statin	71	77	35	76	0.02	1	.89
ACE-I	25	27	21	46	4.71	1	.03
Anticoagulation	4	4	5	11	2.14	1	.14
Family history of AAA	8	9	4	9	0.00	1	1
History of  Thoracic EVAR	2	2	1	2	0.00	1	1
Carotid endarterectomy or stent	5	5	4	9	0.54	1	.47
AAA repair	5	5	1	2	0.78	1	.38
Iliac aneurysm	32	35	28	61	8.49	1	.01
Bilateral iliac aneurysm	18	20	16	35	3.83	1	.05

AAA, Abdominal aortic aneurysm; ACE-I, angiotensin-converting enzyme inhibitor; ASA, American Society of Anesthesiologists; CHF, congestive heart failure; EVAR, endovascular aneurysm repair.

^aThe P value for each variable was calculated by χ² analysis. Statistically significant variables (P < .05) are in bold.

^bSanofi-Aventis, Bridgewater, NJ.

Perioperative and postoperative outcomes and increased LOS after EVAR

Table II reports the perioperative variables that were recorded and analyzed. Performance of concomitant procedures (P < .001), especially renal angioplasty (P = .01) and hypogastric coiling, both intra-operatively (P = .01) and preoperatively (P < .001), resulted in significantly increased LOS. LOS was prolonged in patients requiring intraoperative thromboembolectomy (P = .01). Open vs percutaneous access technique was not a predictor of prolonged LOS in this study; almost 80% of patients in both groups had percutaneous access. A type II endoleak was the most common endoleak, with an incidence of ~20%; however, the presence of an endoleak was not associated with prolonged LOS. It was reassuring to note that the brand of endograft used also did not have any effect on LOS.

Table II.

Intraoperative variables by length of stay (LOS)

	LOS ≤2 days		LOS >2 days
Intraoperative variables	No.	%	No.	%	χ 2	df	Pa
Unfit for open AAA repair	5	5	4	9	0.53	1	.47
Open femoral access	21	23	11	24	0.02	1	.89
Graft
Endurantb	33	36	18	39	0.14	1	.71
Zenithc	22	24	7	15	1.40	1	.24
Excluderd	13	14	5	11	0.29	1	.59
Endurant IIb	9	10	5	11	0.04	1	.84
Ovatione	2	2	3	7	1.66	1	.20
AFXf	2	2	0	0	1.02	1	.32
Other	11	12	8	17	0.76	1	.38
Aortouniiliac graft	3	3	5	11	3.25	1	.07
Hypogastric covered
Intentionally	12	13	11	24	2.61	1	.11
Unintentionally	0	0	1	2	2.02	1	.16
Arterial injury	0	0	1	2	2.01	1	.16
Endoleak
Type I	8	9	6	13	0.64	1	.43
Type II	19	21	10	22	0.02	1	.88
Type III	1	1	1	2	0.25	1	.61
Concomitant procedure	26	28	30	65	17.37	1	<.001
Hypogastric coil
Preoperative	2	2	8	17	10.57	1	<.001
Intraoperative	3	3	7	15	6.52	1	.01
Unplanned graft extension	3	3	5	11	3.25	1	.07
Femoral endarterectomy	4	4	4	9	1.06	1	.30
Femoral-femoral artery bypass	2	2	3	7	1.66	1	.20
Iliofemoral Bypass	0	0	1	2	2.01	1	.16
Thromboembolectomy	0	0	3	7	6.13	1	.01
Iliac angioplasty	5	5	4	9	0.53	1	.47
Iliac stent placement	7	8	7	15	1.95	1	.16
Renal PTA stent	3	3	7	15	6.52	1	.01
Other arterial reconstruction	6	7	7	15	2.72	1	.10
Antibiotic therapy
Start ≤1 hour of operation	90	98	46	100	1.01	1	.31
Stop ≤24 hours postoperation	79	86	40	87	0.03	1	.86
Cephalosporin (1st or 2nd generation)	86	93	45	98	1.20	1	.27

AAA, Abdominal aortic aneurysm; PTA, percutaneous transluminal angioplasty.

^aThe P value for each variable was calculated by χ² analysis. Statistically significant variables (P< .05) are in bold.

^bMedtronic, Minneapolis, Minn.

^cCook, Bloomington, Ind.

^dW. L. Gore and Assoc, Flagstaff, Ariz.

^eTriVascular, Santa Rosa, Calif.

^fEndologix, Irvine, Calif.

Postoperative complications and variables contributing to prolonged LOS can be found in Table III. Patients unable to extubate immediately postoperatively (P = .01), those requiring pressors (P < .001), or returning to the operating room (P = .01) all had significantly prolonged hospital stays. Cardiac and renal injury were a major predictor of increased LOS in patients, with postoperative MI (P = .04), new dysrhythmias (P < .001), and change in renal function (P = .01) all reaching statistical significance. Of note, patients with prolonged LOS were significantly more likely to go home with ACE-I (P = .03) or anticoagulation (P = .02) therapy. Finally, patients prolonged LOS had with higher mortality at 1 month (P = .05) and 12 months (P = .03).

Table III.

Postoperative management of patients based on length of stay (LOS)

Postoperative complications	LOS ≤ 2 days		LOS >2 days
	No.	%	No.	%	χ 2	df	Pa
Immediate extubation	91	99	40	87	9.10	1	.01
Vasopressors required	1	1	11	24	20.13	1	<.001
MI	0	0	2	4	4.06	1	.04
Dysrhythmia	0	0	8	17	16.98	1	<.001
Respiratory	0	0	0	0	0	1	1
Change of renal  function	0	0	3	7	6.13	1	.01
Leg ischemia emboli	0	0	0	0	0	1	1
Bowel ischemia	0	0	0	0	0	1	1
Wound complication	0	0	0	0	0	1	1
Return to operating  room	0	0	3	7	6.13	1	.01
Stroke	0	0	0	0	0	1	1
Discharged on
Aspirin	78	85	37	80	0.42	1	.52
Plavixb	16	17	6	13	0.43	1	.51
Statin	70	76	30	65	1.82	1	.18
β-blocker	51	55	30	65	1.21	1	.27
ACE-I	3	3	6	13	4.81	1	.03
Anticoagulant	1	1	4	9	5.08	1	.02
Mortality
30 days	0	0	2	4	3.89	1	.05
12 months	3	3	6	13	4.81	1	.03

ACE-I, Angiotensin-converting enzyme inhibitor; MI, myocardial infarction.

^aThe P value for each variable was calculated by χ² analysis. Statistically significant variables (P < .05) are in bold.

^bSanofi-Aventis, Bridgewater, NJ.

Table IV reports the continuous variables assessed with the Mann-Whitney U test for statistical significance. Preoperative hemoglobin (P = .01) and the maximum anterior-to-posterior diameter of the AAA (P = .03) were predictors of prolonged LOS, with 75% of patients with an AAA >6.3 cm staying >2 days. Intraoperative factors predicting extended hospital LOS included total procedure time (P < .001), volume of iodinated contrast (P = .05), volume of intraoperative crystalloid (P = .05), and blood transfusions (P < .001). Intensive care unit (ICU) admission was significantly higher in patients with a LOS >2 days (P < .001).

Table IV.

Continuous variables by length of stay (LOS)

Continuous variables	LOS ≤2 days
	Mean	SD	Medium	25%	75%
Height, cm	175.8	8.6	175	170	182.5
Weight, kg	87.3	18.5	84	75.5	98.5
Body mass index, kg/m2	28.2	5.4	26.9	24.8	29.9
Creatinine, mg/dL	1.0	0.4	1	0.8	1.15
Pre-op hemoglobin, g/dL	14.1	1.8	14.2	12.9	15.5
AAA max diameter, mm	56.2	11.3	54	50.5	59.5
Iliac aneurysm diameter, mm	30.9	11.6	29	22	36
Total procedure time, min	154.8	61.1	136	114	194
Graft body diameter, mm	28.5	4.3	28	25	32
Iodinated contrast, mL	153.1	59.6	150	120	190
Crystalloid, mL	1789	738	1750	1300	2000
Estimated blood loss, mL	179	203	100	50	200
RBCs transfused, units	0.1	0.3	0	0	0
Heart rate on arrival to OR	72.3	12.0	71	64	80
Highest intra-op heart rate	86.9	15.0	85	76	98
ICU, days	0.2	0.6	0	0	0
Hospital
Charges, $	100,000	41,000	94,000	77,000	120,000
Cost, $	27,000	10,000	26,000	20,000	32,000
Physician charges	8000	5700	7400	4500	10,000

LOS >2 days
Mean	SD	Medium	25%	75%	Pa
174.2	9.0	175	168	183	.36
83.3	23.6	81	69	91	.08
27.3	6.4	26.6	23.2	29.2	.23
1.1	0.4	1	0.8	1.2	.89
12.9	2.0	13.3	11.5	14.4	.01
59.1	10.7	59.5	53	63	.03
31.9	11.9	31	23	37	.66
216.0	98.8	190.5	144	271	<.001
29.7	4.9	29.5	26	34	.10
182.3	85.4	165	130	220	.05
2161	1006	2000	1500	2800	.05
413	756	200	100	500	.01
1.2	3.1	0	0	1	<.001
71.4	14.4	68	64	78	.51
85.2	15.3	85.5	74	100	.82
2.4	3.9	1	0	3	<.001
180,000	73,000	170,000	120,000	220,000	.01
45,000	19,000	44,000	33,000	56,000	.01
12,000	12,000	8600	5700	15,000	.12

AAA, Abdominal aortic aneurysm; ICU, intensive care unit; OR, operating room; RBCs, red blood cells; SD, standard deviation.

^aP value for each variable calculated by Mann-Whitney U analysis. Statistically significant variables (P < .05) are in bold.

Hospital cost and physician charge data were obtained through the Clinical Data Repository at our institution. Patients with hospital stay >2 days have a statistically significant increase in hospital charges (P = .01), with a mean hospital bill of $180,000 in the prolonged LOS group compared with $100,000 for patients discharged by postoperative day 2. Hospital costs were also significant (P = .01), with a cost of $45,000 for prolonged LOS vs $27,000 in for LOS ≤2 days. Finally, there was no statistical difference in physician charges (P = .09) between the two groups, at $8000 compared with $12,000 for prolonged LOS.

DISCUSSION

The present study found a number of risk factors, including multiple modifiable risk factors, were predicative of an increased LOS. Preoperative characteristics shown to independently predict prolonged LOS were race, ASA score, AAA diameter, severe CHF, ACE-I therapy, and hemoglobin level. Although we are unable to change many of the characteristics of our patient population, this information is important to consider when setting a standard metric for LOS. Intraoperative factors predicting LOS >2 days included procedure time, volume of contrast, volume of crystalloid, hypogastric coiling, performance of a concomitant procedure, and any patient requiring intraoperative thromboembolectomy. Many of these factors are adjustable, requiring strict attention to moderate the risk for increased morbidity and mortality associated with increased LOS. Finally, postoperative complications are good targets for modifiable risk reduction after EVAR. Complications include inability to extubate immediately postoperatively, vasopressor requirements, blood transfusions, ICU admission, return to the operating room, and cardiac and renal dysfunction.

Preoperative patient characteristics that increase LOS can usually not be modified but can help better risk-stratify patients for mortality as well as perhaps predict LOS. Known preoperative predictors before EVAR are especially important for individuals who may need discharge to a skilled nursing facility. In the present study, ASA score was the strongest preoperative predictor of prolonged LOS, with almost half of patients hospitalized >2 days having an ASA of ≥4, whereas >85% of patients discharged by postoperative day 2 were ASA ≤3.

Patients undergoing EVAR are inherently unhealthy; however, that severe CHF was the only comorbidity independently associated with prolonged LOS is surprising.^13-18 These patients were unable to perform any physical activity without discomfort and had symptoms of cardiac insufficiency at rest. Because many of these patients receive ACE-I medications as a cornerstone of medical therapy for CHF, seeing the statistical association between preoperative ACE-I and prolonged LOS is not surprising. It is important to consider stopping medications in the class preoperatively; however, more work needs to be done examining this phenomenon before we make final recommendations. A low hemoglobin level, which was also associated with prolonged LOS, likely serves as a marker of chronic illness.

Procedure details are potentially modifiable and represent opportunities to reduce LOS. Prolonged operative time, increased blood loss, volume of iodinated contrast, and volume of crystalloid resulted in prolonged LOS. These intraoperative variables are likely markers of case complexity but must be taken into consideration when attempting to im prove outcomes. Although there are inherent differences in practice between surgeons that might explain some variability in intraoperative variables, it is important to note that in the present series, there were no differences in LOS for the four surgeons performing EVAR.

Buck et al¹⁹ reports a significant reduction in operative time, postoperative complications, and LOS after percutaneous access compared with open; however, we did not see a difference in LOS for our population. This is likely explained by their high rate of open access, with only 26% of patients undergoing bilateral percutaneous access. Our group is a percutaneous-first center, with almost 80% bilateral percutaneous access for EVAR.

The performance of concomitant procedures was associated with patients staying in the hospital longer. This may be relevant to consider performing secondary procedures, such as coil embolization of internal iliac arteries, large inferior mesenteric arteries, or accessory renal arteries, in a separate stay. We were not surprised to see that patients requiring hypogastric artery coiling stayed longer, regardless of when the procedure was performed, because this is a marker of AAA complexity.¹⁸

A type II endoleak was the most common endoleak, with an incidence of ~20%, comparable to the 19% to 30% reported by other groups, However, the presence of an endoleak was not associated with prolonged LOS.^20-23 It was reassuring to note that the brand of endograft used also did not have any effect on LOS. This confirms what has been demonstrated in most studies.^13-18 Patients with larger AAAs remained hospitalized longer and are known to have higher complication rates.^13,18

Postoperative complications influence LOS and represent the greatest opportunity for improvement.¹⁸ As opposed to preoperative risk factors, which were not predictive of increased LOS, cardiac and renal postoperative complications were some of the strongest predictors of prolonged LOS. This further supports the paradigm that suggests that it is critical to optimize medical management of these comorbidities preoperatively and, importantly, return patients to their medications as soon as possible after their operation. However, patients being discharged with ACE-I and anticoagulation therapy were more likely to have prolonged LOS. Once again, ACE-I therapy is confounded by the increased LOS secondary to prolonged LOS. Patients discharged home on new anticoagulation medication likely sustained an in-hospital complication, such as deep venous thrombosis, pulmonary embolism, atrial fibrillation, or thromboembolism, necessitating prolonged hospitalization.

Acute kidney injury (AKI) after a contrast load in a patient with chronic kidney disease is another modifiable complication after EVAR that is directly related to increased morbidity and mortality.^24-26 It is critical to identify patients with chronic kidney disease preoperatively and optimize them as much as possible and even to pretreat them to reduce the risk of significant AKI.²⁶ Reducing the total volume of contrast and even using alternative contrast agents, such as CO₂, will also help prevent AKI that frequently results in prolonged LOS and increased cost.^24,27,28

Other factors predicting increases in LOS after EVAR included an inability to extubate postoperatively, vasopressor/blood requirement, admission to the ICU, and return to the operating room. However, these are often unpreventable when caring for critically ill patients with complex vascular disease.¹¹

Another important finding in this study is the association between prolonged LOS after EVAR and mortality. Mortality at 30 days and 1 year was significantly higher in the prolonged LOS group, highlighting the long-term effect of periprocedural complications. Prolonged hospitalization leads to increased cost to the health care system and to greater risk for hospital-related morbidity.² Postoperative admission to the ICU is not always necessary; alternatives, such as specialized vascular floors, could offer cost containment while preserving quality outcomes.²⁸ In addition, these patients have a small physiologic window and do not respond well to changes in their hemodynamic state; therefore, hospital-acquired infections, disruption of their complex medication regimens, and invasive procedures place these patients at great risk for increased mortality with prolonged hospitalization.²

The final significant finding in this study is the increase in cost to the health care system for patients with prolonged LOS. We demonstrated significantly higher hospital costs and charges, with no change in physician’s charges. As we move toward bundled payments for procedure-related hospital admissions, reducing the hospital cost is critical because the charges will be fixed regardless of the amount of care a patient requires. It is interesting to note that there was a wide variation in charges identified in this study, with very large standard deviations. This illustrates the ongoing problem with identifying and reducing medical expense in our country: all patients received the same operation, but costs differed greatly.

The limitations of this study include its retrospective nature, the relatively small sample size achieved with a single-center study, and lack of long-term procedure-related complication data >1 year. The Society for Vascular Surgery Vascular Quality Initiative database does an outstanding job capturing perioperative data and 30-day outcomes; however, long-term mortality data were more difficult to obtain through our Clinical Data Repository. It would be very helpful to have more information on procedure-related complications >30 days, including late endoleaks and reoperation, to determine if LOS impacts these factors.

This study provides an important message for surgeons performing EVAR in patients with complex aortic disease. Given the current culture of value-driven health care with a focus on quality and cost, providing the most efficient and effective care to our patients is more important than ever. As we move into future value-based purchasing driven by third-party payers, we will be required to provide the highest quality of care, by standard metrics, at the lowest cost. Although many of the factors predicting prolonged LOS in our population are nonmodifiable, this study demonstrates the importance of identifying patients at risk for postoperative complications and optimizing medical management of these comorbidities.

CONCLUSIONS

This study highlights several modifiable risk factors leading to increased LOS after EVAR, including performance of concomitant procedures, admission to the ICU, and prevention of postoperative renal and cardiac complications. Further, increased LOS was associated with increased charges and costs after EVAR. This study highlights specific areas of focus for improving quality in vascular surgery.